Switching circuit with voltage bias
In one embodiment, an RF impedance matching network for a plasma chamber is disclosed. The matching network includes an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance of the EVC. Each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor. For each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET.
Latest Patents:
The present application is a continuation in part of U.S. patent application Ser. No. 16/722,219, filed Dec. 20, 2019, which is a continuation in part of U.S. patent application Ser. No. 16/673,220, filed Nov. 4, 2019, which is a continuation in part of U.S. patent application Ser. No. 16/667,293, filed Oct. 29, 2019, which is a continuation in part of U.S. patent application Ser. No. 16/654,788, filed Oct. 16, 2019, which is a continuation in part of U.S. patent application Ser. No. 16/415,764, filed May 17, 2019, which is a continuation in part of U.S. patent application Ser. No. 15/816,351, filed Nov. 17, 2017, which is a continuation in part of U.S. patent application Ser. No. 15/450,495, filed Mar. 6, 2017, which is a continuation in part of U.S. patent application Ser. No. 15/196,821, filed Jun. 29, 2016, which claims the benefit of U.S. Provisional Patent Application No. 62/185,998 filed on Jun. 29, 2015.
U.S. patent application Ser. No. 15/450,495 further claims the benefit of U.S. Provisional Patent Application No. 62/303,625, filed Mar. 4, 2016. U.S. patent application Ser. No. 15/816,351 further claims the benefit of U.S. Provisional Patent Application No. 62/424,162, filed Nov. 18, 2016. U.S. patent application Ser. No. 16/654,788 further claims the benefit of U.S. Provisional Patent Application No. 62/751,851, filed Oct. 29, 2018. U.S. patent application Ser. No. 16/667,293 further claims the benefit of U.S. Provisional Patent Application No. 62/753,959, filed Nov. 1, 2018, and U.S. Provisional Patent Application No. 62/767,717, filed Nov. 15, 2018. U.S. patent application Ser. No. 16/673,220 further claims the benefit of U.S. Provisional Patent Application No. 62/754,768, filed Nov. 2, 2018. U.S. patent application Ser. No. 16/722,219 further claims the benefit of U.S. Provisional Patent Application No. 62/784,590, filed Dec. 24, 2018. The present application further claims the benefit of U.S. Provisional Patent Application No. 62/788,269, filed Jan. 4, 2019. The disclosures of the aforementioned priority applications are incorporated herein by reference in their entirety.
BACKGROUNDVariable capacitors are used in many applications, such as matching networks and variable filters. They allow for the precise tuning of frequency and/or impedance in applications needing a dynamic system response, such as in plasma processes. The ability to dynamically change impedance and frequency response provides more flexibility for the applications variable capacitors are used in, and can compensate for variations from unit-to-unit. Some examples of variable capacitors are vacuum variable capacitors (VVCs) and electronically variable capacitors (EVCs).
In electronic circuits, matching networks are used to match the source impedance to the load impedance and vice versa. That is, the source, being of some impedance with a resistive part and a reactive part, will be terminated into the complex conjugate impedance, and the load impedance will be driven by the complex conjugate of its impedance. The complex conjugate is used to eliminate the reactive part of the impedance, leaving only the resistive part, and the resistive part is made equal. This is done so that maximum power transfer can be achieved at the load.
In plasma applications, the load impedance can vary depending on several factors, such as time, power level, pressure, gas flow, chemistry of the gasses, and whether the plasma has been struck. Accordingly, the matching network must be able to automatically vary itself to ensure that the maximum power transfer is achieved. This helps with repeatability in both the depositing and etching.
While the performance of matching networks has improved, there is need for faster and more reliable tuning that meets certain system requirements. For example, there is need to switch an EVC in a manner that increases switching while reducing variability.
BRIEF SUMMARYIn one aspect, an impedance matching network includes a radio frequency (RF) input configured to operably couple to an RF source; an RF output configured to operably couple to a plasma chamber; and an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance of the EVC; wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor; and wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET.
In another aspect, a method of matching an impedance includes coupling an RF input of a matching network to an RF source; coupling an RF output of the matching network to a plasma chamber, wherein the matching network comprises an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit configured to switch in and out the discrete capacitor, wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor, and wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET; and matching an impedance by at least one of the switching circuits of the EVC switching in or out its corresponding discrete capacitor to alter a total capacitance of the EVC.
In another aspect, a semiconductor processing tool includes a plasma chamber configured to deposit a material onto a substrate or etch a material from the substrate; and an impedance matching network operably coupled to the plasma chamber, the matching network comprising an RF input configured to operably couple to an RF source; an RF output operably coupled to the plasma chamber; and an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance of the EVC; wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor; and wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET.
In another aspect, a method of fabricating a semiconductor, the method includes placing a substrate in a plasma chamber configured to deposit a material layer on the substrate or etch a material layer from the substrate; energizing plasma within the plasma chamber by coupling RF power from an RF source to the plasma chamber to perform the deposition or etching; and while energizing the plasma, carrying out an impedance match by an impedance matching network coupled between the plasma chamber and the RF source, the matching network comprising an RF input operably coupled to the RF source; an RF output operably coupled to the plasma chamber; and an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance of the EVC; wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor; and wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET.
In another aspect, an electronically-variable capacitance system includes discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance provided by the variable capacitance system; wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor; and wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET.
The present disclosure will become more fully understood from the detailed description and the accompanying drawings, wherein:
The following description of the preferred embodiment(s) is merely exemplary in nature and is in no way intended to limit the invention or inventions. The description of illustrative embodiments is intended to be read in connection with the accompanying drawings, which are to be considered part of the entire written description. In the description of the exemplary embodiments disclosed herein, any reference to direction or orientation is merely intended for convenience of description and is not intended in any way to limit the scope of the present invention. The discussion herein describes and illustrates some possible non-limiting combinations of features that may exist alone or in other combinations of features. Furthermore, as used herein, the term “or” is to be interpreted as a logical operator that results in true whenever one or more of its operands are true. Furthermore, as used herein, the phrase “based on” is to be interpreted as meaning “based at least in part on,” and therefore is not limited to an interpretation of “based entirely on.”
Features of the present invention may be implemented in software, hardware, firmware, or combinations thereof. The computer programs described herein are not limited to any particular embodiment, and may be implemented in an operating system, application program, foreground or background processes, driver, or any combination thereof. The computer programs may be executed on a single computer or server processor or multiple computer or server processors.
Processors described herein may be any central processing unit (CPU), microprocessor, micro-controller, computational, or programmable device or circuit configured for executing computer program instructions (e.g., code). Various processors may be embodied in computer and/or server hardware of any suitable type (e.g., desktop, laptop, notebook, tablets, cellular phones, etc.) and may include all the usual ancillary components necessary to form a functional data processing device including without limitation a bus, software and data storage such as volatile and non-volatile memory, input/output devices, graphical user interfaces (GUIs), removable data storage, and wired and/or wireless communication interface devices including Wi-Fi, Bluetooth, LAN, etc.
Computer-executable instructions or programs (e.g., software or code) and data described herein may be programmed into and tangibly embodied in a non-transitory computer-readable medium that is accessible to and retrievable by a respective processor as described herein which configures and directs the processor to perform the desired functions and processes by executing the instructions encoded in the medium. A device embodying a programmable processor configured to such non-transitory computer-executable instructions or programs may be referred to as a “programmable device”, or “device”, and multiple programmable devices in mutual communication may be referred to as a “programmable system.” It should be noted that non-transitory “computer-readable medium” as described herein may include, without limitation, any suitable volatile or non-volatile memory including random access memory (RAM) and various types thereof, read-only memory (ROM) and various types thereof, USB flash memory, and magnetic or optical data storage devices (e.g., internal/external hard disks, floppy discs, magnetic tape CD-ROM, DVD-ROM, optical disk, ZIP™ drive, Blu-ray disk, and others), which may be written to and/or read by a processor operably connected to the medium.
In certain embodiments, the present invention may be embodied in the form of computer-implemented processes and apparatuses such as processor-based data processing and communication systems or computer systems for practicing those processes. The present invention may also be embodied in the form of software or computer program code embodied in a non-transitory computer-readable storage medium, which when loaded into and executed by the data processing and communications systems or computer systems, the computer program code segments configure the processor to create specific logic circuits configured for implementing the processes.
Ranges are used as shorthand for describing each and every value that is within the range. Any value within the range can be selected as the terminus of the range. In addition, all references cited herein are hereby incorporated by referenced in their entireties. In the event of a conflict in a definition in the present disclosure and that of a cited reference, the present disclosure controls.
In the following description, where circuits are shown and described, one of skill in the art will recognize that, for the sake of clarity, not all peripheral circuits or components are shown in the figures or described in the description. Further, the terms “couple” and “operably couple” can refer to a direct or indirect coupling of two components of a circuit.
Voltage Reduction Circuit
Referring now to
In the exemplified embodiment, the system 10 includes a radio frequency (RF) source 30 having a substantially fixed output impedance Rsource (e.g., 50 ohms). The RF source 30 generates an RF signal that is received at the input 101 of the matching network 100. The RF source 30 is also operably coupled to chassis ground GND. The RF source 30 may be an RF generator of a type that is well-known in the art to generate an RF signal at an appropriate frequency and power for the process performed within the load 20. The RF source 30 may be electrically connected to the RF input 101 of the impedance matching network 100 using a coaxial cable or similar means, which for impedance matching purposes may have the same fixed (or substantially fixed) impedance as the RF source 30.
The system 10 further includes a load. In the exemplified embodiment, the load is a plasma chamber 20 for manufacturing a semiconductor. The semiconductor device can be a microprocessor, a memory chip, or another type of integrated circuit or device.
As is known in the art, the plasma within a plasma chamber 20 typically undergoes certain fluctuations outside of operational control so that the impedance presented by the plasma chamber 20 is a variable impedance. Since the variable impedance of the plasma chamber 20 cannot be fully controlled, an impedance matching network may be used to create an impedance match between the plasma chamber 20 and the RF source 30. In other embodiments, the load can be any load of variable impedance that can utilize a matching network.
The plasma chamber 20 can include a first electrode 22 and a second electrode 26, and in processes that are well known in the art, the first and second electrodes, in conjunction with appropriate control systems (not shown) and the plasma in the plasma chamber 120, enable one or both of deposition of materials onto a substrate 24 and etching of materials from the substrate 24. The plasma chamber 20 can receive an RF signal from the output 102 of the matching network 100 and thereby receive RF power from the RF source 30 to energize plasma within the plasma chamber 20 to perform the deposition or etching.
The matching network 100 can consist of a single module within a single housing designed for electrical connection to the RF source 30 and plasma chamber 20. In other embodiments, the components of the matching network 100 can be located in different housings, some components can be outside of the housing, and/or some components can share a housing with a component outside the matching network 100.
The matching network 100 provides impedance matching for the RF source 30 and the plasma chamber 20. The matching network 100 is operably coupled between the RF source 30 and the plasma chamber 20. The matching network 100 includes an input 101 configured to operably couple to the RF source 30, and an output 102 configured to operably couple to the plasma chamber 20. The matching network 100 further includes a first variable capacitor C1 and a second variable capacitor C2. In a preferred embodiment, the variable capacitors C1, C2 are EVCs, though in other embodiments, other types of variable capacitors can be used, such as VVCs. EVCs may use switches to add or remove the discrete capacitors, such as an MLCC (multi-layer ceramic capacitor), that form the EVC. The capacitor-switch circuit may be placed in parallel with other capacitor-switch circuits. The parallel circuits allow the discrete capacitors to be simply added or subtracted in the circuit, depending on how many switches are opened or closed. In the case where all the switches are open, the EVC will be at its lowest capacitance value. In the case where they are all closed, the EVC will be at its highest capacitance value.
In this first embodiment, the matching network 100 is a pi network. The first variable capacitor C1 forms part of a first shunt S1 parallel to the RF source 30, and the second variable capacitor C2 forms part of a second shunt S2 separate from the first shunt S1. Put differently, the first variable capacitor C1 is parallel to the input 101, and the second variable capacitor C2 is parallel to the output 102. Further, a first inductor L1 is located between the first shunt S1 and the second shunt S2. In other embodiments, a second inductor L2 can be located between the second shunt S2 and the output 102.
The first variable capacitor C1 has a first capacitance, and the second variable capacitor C2 has a second capacitance. The first capacitance and the second capacitance are configured to be altered to create an impedance match at the input. As will be discussed further herein, however, the invention is not limited to pi matching networks, as other types of matching networks can be utilized.
To reduce a voltage on the second variable capacitor C2, the matching network 100 further includes a third capacitor C3 in series with the second variable capacitor C2. Components or nodes are said to be “in series” if the same current flows through each. In the exemplified embodiment, the third capacitor C3 forms part of the second shunt S2, though the invention is not so limited. In other embodiments, the third capacitor C3 can be at different locations, provided the third capacitor C3 is positioned to reduce a voltage on the second variable capacitor C2 (the reduced voltage being, for example, an alternating current or radio frequency voltage). For example, the positions of C2 and C3 in
In the exemplified embodiment, a fourth capacitor C4 is included. The fourth capacitor C4 is parallel to the second shunt S2 and helps to offset the total capacitance. In other embodiments, the fourth capacitor C4 can be omitted.
In the embodiment discussed below, the values of the additional fixed capacitor C3 and variable capacitors C2 (see
The voltage drop VDrop across the variable capacitor C2 (see
If C2Max=C3, then the formula can be simplified as below, where C2Max=C3=C.
As a result, VDrop is equal to half of the voltage that was originally capacitor C2 (VC2) when C3 was not included.
Continuing with this example, the next step is to find the maximum capacitance required for the variable and fixed capacitors. In this case, the total series capacitance CVar is equal to the maximum capacitance of the original variable capacitor C2. The capacitance CVar can be calculated by the following equation:
If C2Max=C3=C, the equation can be modified as follows:
C is then solved for as follows:
C=2*C VarMax
The minimum value for variable capacitor C2, C2Min, can be found by using the previously calculated value for C3 and replacing the CVarMax with the minimum capacitance, CVarMin, as in the following equations:
It can also be seen, however, that the first (usable) region 702 has gaps representing areas where a perfect impedance match is not provided. This can be a result of adding capacitor C3 to reduce the voltage, which increases the gap between the quantized states of the variable capacitor when approaching C2Min and decreased the spacing when approaching C2Max.
The addition of a third variable or non-variable capacitor, to help further reduce VDrop, can change the capacitor range of the variable capacitor combination C2. To address this, a variable capacitor such as an EVC can be easily modified to adjust the capacitor range. The third capacitor can also change the step sizes and make them nonlinear. In certain embodiments, a more uniform distribution can be provided by using a nonlinear variable capacitor or multiple variable capacitors in series.
In other embodiments, transmission lines (which can comprise microstrips, coaxial cable, a wave guide, or any other conductive medium) can be used to rotate the impedance of the matching network on the Smith chart. The length of the transmission line at a certain frequency determines the amount of rotation. The longer the transmission line, the more rotation there will be on the Smith chart. A quarter wavelength (λ/4) transmission line (which can be calculated using the operating frequency and the property of the dielectric material) will have a 180° clockwise rotation on the Smith chart, a half wavelength (λ/2) transmission line will have a 360° clockwise rotation on the Smith chart, an eighth wavelength (λ/8) would be equal to 45°, and so on.
If the matching network 1000 uses only quarter wave lines, or something that would ultimately give a 90° phase shift [(λ/4)+N*(λ/2)], and there are the three capacitors C101, C102, C103 in shunt (together with transmission lines TL1 and TL2), as shown in
The foregoing embodiments provide several advantages. The embodiments disclose a matching network that can more effectively handle high voltages generated in a network. Further, the embodiments avoid or minimize the need for increased component sizes (as typically required for a VVC) or increased numbers of peripheral components (as typically required with an EVC). Further, the embodiments provide a solution that has a lower cost than previous methods of addressing high voltages in a matching network. As shown herein, the embodiments can increase the usable range of a matching network without sacrificing the impedance range, using a more expensive, larger, higher voltage component, or adding more peripheral components to meet the voltage requirements.
As discussed above, an EVC is a type of variable capacitor that can use multiple switches, each used to create an open or short circuit, with individual series capacitors to change the capacitance of the variable capacitor. The switches can be mechanical (such as relays) or solid state (such as PIN diodes, transistors, or other switching devices). The following is a discussion of various methods for setting up an EVC or other variable capacitor to provide varying capacitances.
In an accumulative setup of an EVC, the approach to linearly increase the capacitor value from the minimum starting point (where all switches are open) is to incrementally increase the number of fine tune capacitors that are switched into the circuit. Once the maximum number of fine tune capacitors is switched into circuit, a coarse tune capacitor is switch in, and the fine tune capacitors are switched out. The process starts over with increasing the number of fine tune capacitors that are switched into circuit, until all fine and coarse tune capacitors are switched in. In this setup, all of the fine tune capacitors have the same or a substantially similar value, and all the coarse tune capacitors have the same or a substantially similar value. Further, the capacitance value of one coarse tune capacitor about equals the combined capacitance value of all fine tune capacitors plus an additional fine tune capacitor into the circuit, thus enabling a linear increase in capacitance.
An example of this in an ideal setting would be if the fine tune capacitors were equal to 1 pF, and the coarse tune capacitors were equal to 10 pF. In this ideal setup, when all switches are open, the capacitance is equal to 0 pF. When the first switch is closed, there is 1 pF in the circuit. When the second switch is closed there is 2 pF in the circuit, and so on, until nine fine tune switches are closed, giving 9 pF. Then, the first 10 pF capacitor is switched into circuit and the nine fine tune switches are opened, giving a total capacitance of 10 pF. The fine tune capacitors are then switched into circuit from 11 pF to 19 pF. Another coarse tune capacitor can then be switched into circuit and all fine tune capacitors can be switched out of circuit giving 20 pF. This process can be repeated until the desired capacitance is reached.
This can also be taken one step further. Using the previous example, having nine 1 pF capacitors and also nine 10 pF capacitors, the variable capacitor circuit can have even larger values, 100 pF, to switch in and out of circuit. This would allow the previous capacitor array to go up to 99 pF, and then the 100 pF capacitor can be used for the next increment. This can be repeated further using larger increments, and can also be used with any counting system.
An alternative capacitor setup is referred to herein as a binary weighted setup. In the binary weighted setup, the capacitor values will all be different. The first value is equal to the minimum desired change in capacitance. Then each successive capacitor value is increased to double the change in capacitance from the previous up until the maximum desired capacitor value, when all capacitors are switched in.
In one example (that assumes there are no parasitic capacitances), the lowest capacitance capacitor would be a 1 pF capacitor, followed by 2 pF, 4 pF, and so on. When all switches are open, the value is 0 pF. When the 1 pF capacitor is switched in, the EVC total capacitance value is 1 pF. Then the 1 pF capacitor is switched out of circuit and the 2 pF capacitor is switched in, causing a total capacitance of 2 pF. When 3 pF is needed, the 1 pF and the 2 pF capacitors are switched in. For 4 pF, the 1 and 2 pF capacitors are switched out of circuit and the 4 pF capacitor is switched into circuit. This can be repeated adding 1 pF, 2 pF, and 4 pF together in different combinations in the circuit, creating values of 5 pF, 6 pF and 7 pF.
In the embodiment of
The binary weighted setup can result in using far less capacitors to switch in and out of circuit to achieve the same or better resolution and range. A potential problem with this setup, however, is that, once the capacitor reaches a certain value, the voltage and/or current on that particular capacitor or the current on the switch can be higher than the specification allows for. This forces the EVC to use multiple capacitors in parallel for each switch of lower value.
Another potential disadvantage of the binary weighted setup is that it is difficult to achieve a consistent step size throughout the range. The above capacitor values for the binary setup give an average step size of 2.32 pF, compared to the accumulative method, which has an average step size of 2.72 pF. But the minimum and maximum step for the binary weighted setup is 1.51 pF and 7.51 pF, respectively, while the accumulative setup's minimum and maximum are only 2.4 pF and 2.75 pF.
With higher value capacitors, this can be further complicated with finding a value that does not overshoot multiple steps. Also, part-to-part tolerances being greater than the minimum step size can further increase the gaps. A 300 pF capacitor with a ±5% tolerance can have up to 15 pF of extra capacitance. The delta capacitance of the three least significant binary weighted capacitors total 15.44 pF. So, these values are completely overstepped, and linearity is lost.
One modification to the binary weighted setup is to have the larger capacitor values rounded down to the next standard value, for example 3.0 pF, 5.1 pF, 9.1 pF, 18 pF, 36 pF, 68 pF, 130 pF, 240 pF. Doing this would create some overlap in capacitor value where there would be a drop in capacitance when switching in the new larger value and switching out the previous smaller values. For example, the values 3 pF through 36 pF would combine to equal 71.2 pF, but the next step is 68 pF, a drop of 3.2 pF. This problem can be avoided, however, because the EVC does not need to go sequentially through each step, but instead can use software to lookup the next known capacitor position to switch to it directly.
The switches 60 can be coupled to switch driver circuits 80 for driving the switches on and off. The variable capacitance system 55 can further include a control unit 85 operably coupled to the driver circuits 80 for instructing the driver circuits 80 to switch one or more of the switches 60, and thereby turn one or more of the capacitors 77 on or off. In one embodiment, the control unit 85 can form part of a control unit that controls variable capacitor, such as a control unit that instruct the variable capacitors of a matching network to change capacitances to achieve an impedance match.
In the exemplified embodiment, the first capacitors 50 are fine tune capacitors using a method similar to the binary method discussed above. Thus, the fine tune capacitors 50 can have capacitances increasing by a factor of about two, where “about two” refers to a value of 1.5 to 2.5. In an ideal example where there are no parasitic capacitances, the fine tune capacitors could increase by a factor of exactly two (e.g., 1 pF, 2 pF, 4 pF, 8 pF).
But in real world applications, parasitic capacitances, such as those provided by the switches 60, are another factor that must be considered in choosing the capacitance values of the fine tune capacitors 50. Thus, while a first capacitor may have a value of 1 pF, and the corresponding capacitor-switch pair may thus provide 1 pF to a total capacitance of the variable capacitor when the capacitor's corresponding switch is closed, when the switch is open, the open switch may have a parasitic capacitance of, for example, 1 pF. Thus, when the switch is open, there are essentially two 1 pF capacitances in series, which is equivalent to 0.5 pF. Thus, when the first fine tune capacitor switch switches from open to close, the change in the capacitance contributed to the variable capacitor by this capacitor-switch pair is from 0.5 pF (open) to 1 pF (closed), for a change of 0.5 pF. These changes in capacitance caused by parasitic capacitances must be taken into consideration in choosing capacitor values to ensure that the target step size (e.g., 0.5 pF) for the total capacitance can be achieved.
Returning to the previous example, if an EVC had four fine capacitors, and each capacitor switch had a parasitic capacitance of 1 pF, and a step size of 0.5 pF was desired, the fine capacitors could be 1 pF, 1.6 pF, 2.7 pF, and 4.7 pF. As discussed, the first fine capacitor (1 pF) would cause a 0.5 pF change to the total capacitance when switched in. The second fine tune capacitor (1.6 pF) and its switch would provide 0.6 pF when open and 1.6 pF when closed, thus causing a change in the total capacitance of about 1 pF when switched in. The third fine tune capacitor (2.7 pF) would cause a change in the total capacitance of about 2 pF when switched in, and the fourth fine tune capacitor (4.8 pF) would cause a change in the total capacitance of about 4 pF when switched in. Thus, the changes to the total capacitance caused by the switching in of each of the four first tune capacitors would be 0.5 pF, 1 pF, 2 pF, and 4 pF, respectively. Thus, the changes caused by the switching in of each of these capacitors increases by a factor of two. It is understood that the invention is not limited to these values. Other capacitor values (or switches with other parasitic capacitances) can be used such that the changes caused increase by a factor of about two. For example, the 4.8 pF capacitor of the above example could be replaced with a standard 4.7 pF capacitor. Further, other capacitance values can be used to achieve other step sizes. The foregoing considerations regarding parasitic capacitances can equally apply to the binary setup discussed above.
The second capacitors 70, by contrast, are coarse tune capacitors using a method similar to the accumulative method discussed above. Thus, the second capacitors can have a substantially similar capacitance. Capacitors are considered to have substantially similar capacitances if, of the capacitors in question, no capacitance is 15 percent (15%) greater than or less than another capacitance. Alternatively, the capacitors can be chosen such that there are no gaps in total capacitance greater than the minimum step size needed for the given application.
The first (fine) capacitors 50 can increase their value (or the value by which they change the total capacitance) in a binary fashion, and thus by a factor of about two, up to the first coarse position. When all of the fine capacitors 50 are switched into circuit, the first coarse capacitor 71 can be switched in, and all the fine capacitors 50 are switched out. Then the fine capacitors 50 can be switched in and out until they are all switched into circuit. The next step would be to add another coarse tune capacitor 72. It is understood, however, that the EVC does not need to go sequentially through each step to achieve a desired total capacitance, but instead can use software to lookup the next known capacitor position to switch to it directly.
In one embodiment, there are four fine capacitors 50. The first fine capacitor 51 has a capacitance of 3.0 pF, the second fine capacitor 52 has a capacitance of 5.1 pF, the third fine capacitor 53 has a capacitance of 9.1 pF, and the fourth fine capacitor has a capacitance of 18 pF. Further, there are four coarse tune capacitors 70 having capacitances of 36 pF each. Thus, in this embodiment, the total combined capacitance of the fine capacitors (35.2 pF) is substantially similar to the individual capacitances of the coarse capacitors (36 pF). It also follows that the capacitance of each of the coarse capacitors is greater than a greatest individual capacitance (18 pF) of the fine capacitors.
In this embodiment, there will be 208 unique capacitor values. With parasitics, the minimum total capacitance is 10.25 pF and the maximum total capacitance is 467.2 pF. The range is less than 1 pF less than the accumulative method, but with an increase in unique points. The minimum step size is 1.51 pF, the maximum is 2.54 pF and the average is 2.21 pF. Thus, the results of the setups discussed are as follows:
The partial binary method provides multiple advantages. First, the current on each capacitor will not be over its rating. The maximum current and the current rating will be the same for all coarse capacitors, because they will be the same value. With the fine steps, all of the capacitor values have a higher ratio of current rating to maximum current. Therefore, no issues should arise.
Further, the partial binary approach avoids large gaps in capacitance steps. Further, less capacitors are needed to have the same range, while the number of unique values can potentially be increased. With less capacitors, the EVC will need less switches, causing the EVC to take up less area. Further, less capacitors will require less hardware to control the switches.
Binary with overlap can also be implemented in this setup to avoid any issues with part tolerance if required. Thus, the coarse capacitor values could be reduced in capacitance. It is further understood that, while the exemplified embodiment uses four first capacitors 50 and four second capacitors 70, other numbers of capacitors can be used. Also, other capacitor values can be used.
It is understood, however, that the EVC does not need to go sequentially through each step, but instead can use software to lookup the next known capacitor position to switch to it directly. It is further understood that a desired total capacitance can be achieved by having switched on a minimal number of capacitors of the plurality of capacitors.
In another embodiment, the variable capacitor can for part of a method of manufacturing a semiconductor, such as the system displayed shown in
The switches 60-1 can be coupled to switch driver circuits 80-1 for driving the switches on and off. The variable capacitance system 55-1 can further include a control unit 85-1 operably coupled to the driver circuits 80-1 for instructing the driver circuits 80-1 to switch one or more of the switches 60-1, and thereby turn one or more of the capacitors 77-1 on or off. In one embodiment, the control unit 85-1 can form part of a control unit for a matching network that controls the capacitances of one or more variable capacitors of the matching network to achieve an impedance match. In the exemplified embodiment, the control unit 85-1 (sometimes referred to as “control circuit”) is configured to (a) determine which of the coarse capacitors and the fine capacitors to have switched in to achieve an impedance match and (b) cause the determined coarse and fine capacitors to be switched in. This can be based on a determination, by the control unit or otherwise, of the variable impedance of the plasma chamber. The invention is not so limited, however, as the determination of the capacitors to switch in can be based on other factors, such as an input impedance at the input of the matching network.
In the exemplified embodiment, the fine capacitors 50-1 have capacitances increasing by a factor of about two, where “about two” refers to a value of 1.5 to 2.5, though the invention is not so limited and the fine capacitors can increase in value in another manner. In an ideal example where there are no parasitic capacitances, the fine tune capacitors could increase by a factor of exactly two (e.g., 1 pF, 2 pF, 4 pF, 8 pF, etc.). But as discussed above, in real world applications parasitic capacitances, such as those provided by the switches 60-1, are another factor that must be considered in choosing the capacitance values of the fine tune capacitors 50-1. In the exemplified embodiment, the fine capacitors 50-1 have values of 47 pF, 91 pF, 180 pF, 390 pF, 750 pF, 1500 pF, though the invention is not limited to these values or this number of capacitors.
According to the restricted partial binary setup, the coarse capacitors are made up of first coarse capacitors 70-1 each having a substantially similar first coarse capacitance, and second coarse capacitors 71-1 each having a substantially similar second coarse capacitance. Capacitors are considered to have substantially similar capacitances if, of the capacitors in question, no capacitance is 15 percent (15%) greater than or less than another capacitance. In the exemplified embodiment, there are 6 first coarse capacitors 70-1 each having a capacitance of 1000 pF, and 12 second coarse capacitors 71-1 each having a capacitance of 3000 pF. Thus, in the exemplified embodiment, one of the fine capacitors (the 1500 pF fine capacitor) has a capacitance greater than the first coarse capacitance of 1000 pF. In other embodiments, more than one of the fine capacitors can have a capacitance greater than the first coarse capacitance. Further, in other embodiments other values and other numbers of coarse capacitors can be used.
Each capacitor of the plurality of capacitors 77-1 provides a change to a total capacitance of the variable capacitor 75-1 when the capacitor is switched in. To gradually increase the total capacitance of the variable capacitor 75-1, the control unit 85-1 can successively switch in, in a predetermined order, each of the first coarse capacitors 70-1, followed by each of the second coarse capacitors 71-1. As for the fine capacitors 50-1, the control unit restricts which fine capacitors can be switched in. That is, it only switches in the fine capacitors 50-1 whose capacitance is less than a capacitance of a next coarse capacitor of the coarse capacitors predetermined to be switched in next.
As discussed above, in the exemplified embodiment, the fine capacitors 50-1 have capacitances substantially equal to 47 pF, 91 pF, 180 pF, 390 pF, 750 pF, 1500 pF; the first coarse capacitors 70-1 comprise six capacitors having capacitances substantially equal to 1000 pF; and the second coarse capacitors 71-1 comprise twelve capacitors having capacitances substantially equal to 3000 pF. To gradually increase capacitance, the fine capacitors can be switched into the circuit in a binary fashion as described above except for the 1500 pF fine capacitor, which is restricted from switching in until all the first coarse capacitors are switched in. When all the 1000 pF first coarse capacitors are switched in, the next coarse capacitor to be switched in is a 3000 pF second coarse capacitor. Thus, once all the 1000 pF first coarse capacitors are switched in, the 1500 pF fine capacitor is able to switch in with the rest of the fine capacitors.
Table 3 below shows the first 167 positions (“Pos.”) for a variable capacitor using the restricted partial binary setup. The total capacitance (“Total Cap.”) for each position is shown, along with the fine capacitors (F1-F6) and first coarse capacitors (C1-C6) switched in for a given position. As can be seen, although the fine capacitors switch in in a somewhat typical binary fashion, the 1500 pF fine capacitor is not able to switch in until position 165, when all of the first coarse capacitors have been switched in.
It is understood that the variable capacitor 75-1 does not need to go sequentially through each step to achieve a desired total capacitance, but instead can use software to lookup the desired capacitor position to switch to it directly. It is further understood that while the exemplified embodiment uses two sets of coarse capacitors (first coarse capacitors and second coarse capacitors) in other embodiments more or less sets of coarse capacitors can be used. For example, three sets of course capacitors could be used, each with a different capacitance value. It is further understood that the variable capacitance system 55-1 can form part of any one of the impedance matching networks discussed above. For example, a matching network may include an input configured to operably couple to an RF source, an output configured to operably couple to a plasma chamber for manufacturing a semiconductor, and the variable capacitance system 55-1 of
The restricted partial binary setup described above provides multiple advantages. For example, by restricting one or more fine capacitors from switching in, the setup avoids overlap, that is, instances where there is more than one solution for a given impedance value. Further, the restricted partial binary setup allows the variable capacitor to provide a large range of capacitance values by allowing for the use of a high percentage of coarse capacitors. Further, the setup avoids large gaps in capacitance values. Further, less capacitors are needed, thus requiring less switches and causing the variable capacitor to take up less area.
In a matching network, a PIN diode may be used as an RF switch for each discrete capacitor of an EVC.
In the exemplified embodiment of
This choke design may have drawbacks. The main problem is that it may affect the overall switching speed, where large voltages can be generated across the inductor. As shown in Equation 1 (voltage across an inductor), the voltage is equal to the inductance times the rate of change in current.
The current comes from the Bias and therefore is fixed. If the inductance of the choke needs to be large, then the bias being applied must gradually increase and decrease. As an example, if the inductance is 60 uH, the bias current is 0.5 Amps, and the switching speed is 10 ns, the voltage generated across the inductor would be 3000 Volts. This could damage components on the driver circuit. This voltage also oscillates and will cause the PIN diode to turn ON and OFF rapidly. If the switching speed is changed to 50 us, this voltage drops to only 0.6 Volts. To drop the inductance would also have a similar effect, but the isolation required prevents this.
The exemplified matching network 90A further includes a filtering capacitor 62A that is used for filtering to help block the RF from the driver circuit 66A. This filtering capacitance is typically a large value to give a low impedance to ground for the RF signal. This large capacitance, however, may add stress to the driver circuit 66A circuitry by creating additional power dissipation. The filtering capacitor 62A and choke inductor 64A can together be considered a filter.
The dissipation is difficult to calculate because of the complexity of the circuit, especially with the slow, nonlinear transition times of the driver circuit 66A. We do know the energy stored in the discrete capacitor 78 and the filtering capacitor 62A and the amount of time that it takes to fully discharge. Thus, the power dissipated per charge and discharge can be roughly calculated for the full system. Equation 2 may be used to calculate the energy stored in the discrete capacitor 78.
Equation 3 may be used to calculate power dissipation from the energy dissipated over time.
If Equation 2 is substituted into Equation 3, the total power dissipated in the system from the discrete capacitor 78 charge/discharge can be calculated as seen in in Equation 4.
Accordingly, the total power dissipation per switching circuit 61A can be calculated. To continue with the previous example, the switching speed is 50 us. If the HVDC is set to 1650V and the total capacitance is 3400 pF, then the total power dissipated in the system is 93 W. This dissipation includes losses in the driver circuit switch, the power supply, the bias resistor, and anything else in series with the capacitor while it either charges or discharges.
Switching Circuit Utilizing Two PIN Diodes in Parallel
In the exemplified embodiment of
As shown in
It will be noted that the first diode 65 has an anode and a cathode, and the second diode 67 has an anode and a cathode. In the exemplified embodiment, the anode of the first diode 65 is operably coupled to a first terminal of the discrete capacitor 78, the cathode of the second diode 67 is operably coupled to the first terminal of the discrete capacitor 78, and the anode is operably coupled to ground. Further, the cathode of the first diode 65 is coupled to the filtering capacitor 62B, the first diode 65 and the filtering capacitor 62B being parallel to the second diode 67. Further, the cathode of the first diode is coupled to the choke 64B, and the choke 64B is coupled to a driver circuit 66B. The invention, however, is not so limited, as other arrangements or components may be utilized. For example, in an alternative embodiment, the diodes' orientation is switched such that, while still in an opposing orientation, the cathode of the first diode is operably coupled to the first terminal of the discrete capacitor, and the anode of the second diode is operably coupled to the first terminal of the discrete capacitor. Further, while the exemplified embodiments use PIN diodes, the invention is not so limited, as other types of diodes may be utilized. For example, in an alternative embodiment, the second diode 67 is a PIN diode but the first diode is a different type of commonly available diode.
The following provides a comparison of the first switching circuit 61A (
For the ON case, each diode 65, 67 has a resistance of 150 mΩ. The discrete capacitor 78 is where the RF voltage will drop. To achieve the required attenuation, the discrete capacitor 78 needs to be 2780 pF or less. This is a very large value, and the match would easily exceed the current rating of the capacitor or diode. Typically, the maximum capacitance used for matching at 13 MHz is around 100 pF. So, this is a nonissue.
For the OFF case, if each diode 65, 67 is assumed to have 2.5 pF of capacitance. To achieve 45 dB of attenuation, the filtering capacitor 62B needs a minimum value of 442 pF. The next standard value would be 470 pF. This gives an attenuation of 45.5 dB.
The power dissipation can now be calculated as before. With this reduced capacitance value, the dissipation goes down to 12.8 W. If the maximum acceptable power dissipation is 93 W, as before, then the switching transition time can be decreased from 50 us to 6.9 us, an 86% reduction. One could go a step further in reducing the capacitance by replacing the filtering capacitor 62B with a series LC resonator. This would create a very low impedance while also reducing the capacitance seen by the driver circuit 66B. Care should be taken, as half of the RF current from the discrete capacitor 78 will now flow through the LC resonator, which can be significant and could produce large voltages across the LC resonator's inductor and capacitor.
The choke 64B (or more advanced filter topology) is not needed, but provides additional isolation to the driver circuit 66B. Since this inductance value will be low, the voltage generated from it is negligible. An additional filtering capacitor may also be placed on the node where the choke meets the driver circuit. This would add to the total capacitance that is seen from the driver, and therefore would increase the dissipation seen from the driver circuit 66B. The other filtering capacitor may be reduced to help reduce this.
The second switching circuit 61B provides many advantages. A major benefit of the second switching circuit 61B is that the RF voltage is significantly dropped before the choke 64B. This means that the choke 64B can be drastically reduced or even eliminated. The high voltage ringing may be eliminated, and therefore the choke 64B may no longer be the limiting factor on switching speed. Further, the filtering capacitor 62B of the second switching circuit 61B may be significantly smaller. Thus, the power dissipated from the switching is reduced, allowing either the switching speed to be increased or the frequency of switching to be increased, or some combination of both. This could also increase the reliability of the driver circuit 66B as the power dissipation is considerably reduced. Using two PIN diodes 65, 67 in parallel per discrete capacitor 78 allows for more current per channel. The current will not be split evenly, unless the filtering capacitor 62B is replaced with an LC resonator, and therefore the maximum current will be doubled.
It is noted that, as a result of using two PIN diodes, the HVDC is dropped across twice as many PIN diodes. To have the same blocking voltage, the HVDC voltage must be doubled. Further, using two PIN diodes will require an adjustment to the Bias. If a dual PIN diode switch was originally used, then there will now be four diodes in series. If there is a 0.5 Volt drop per diode, the total drop would be 2 Volts instead of one. This may require the Bias voltage to be increased, which could require it to have a higher power dissipation, or require multiple bias supplies.
It is further noted that the switching circuits discussed above may be used as part of a method of matching impedance, or a method of manufacturing a semiconductor, where the switching circuits are used to switch in or out discrete capacitors of an EVC to thereby cause an impedance match. Further, a matching network using one or more of the switching circuits discussed above may for part of a semiconductor processing tool (such as tool 91A or 91B), the semiconductor processing tool further comprising a plasma chamber (such as chamber 21).
EVC-based impedance matching networks can utilize frequency tuning, where the matching network has full control of the frequency of the generator. This may offer large cost advantages, but may also have disadvantages. The two major disadvantages are a reduction in efficiency and/or a narrow tuning range, which are inversely related to each other. If the tuning range covers a large area, the efficiency will be reduced, and if the efficiency needs to be high, the tuning range will be limited. These disadvantages are caused by the frequency tuning elements in the match, particularly the inductor of the LC series network. This network creates a large change in impedance as frequency is varied, more than what either component could generate alone.
The efficiency is another story. As an example, if a load of 0.5-j50 were placed at the output of the circuit, and other ideal components were added to the matching network, when tuned, the efficiency of circuit 17 would be 60.8%, while the efficiency of circuit 19 would be 87.0%. This means that circuit 19 would deliver 43% more power than circuit 17. A similar drop in efficiency would occur if the frequency tuning elements were in shunt. The difference in efficiency is caused by the difference in ESR (equivalent series resistance). At j50Ω, circuit 17 has an ESRC of 2 mΩ for capacitor 17C, and an ESRL of 320 mΩ for inductor 17L, for a total resistance of 322 mΩ. By contrast, circuit 19 has an ESRC of 2 mΩ for capacitor 19C, and an ESRL of 73 mΩ for inductor 19L, for a total resistance of only 75 mΩ. Even though the quality factor of both inductors is well over 1000, when the series capacitor is added, the apparent inductance is reduced. Both apparent inductances are equal, but circuit 17 has about 3.3× more ESR. The minimum loss would occur when the capacitor is completely removed from the circuit, but this would give the smallest possible tuning range.
It is known that one way to increase the range of a matching network without significantly affecting the efficiency is to utilize a variable capacitor. The following embodiment will use a vacuum variable capacitor (VVC). In other embodiments, the VVC can be replaced with another mechanically variable capacitor (MVC). An MVC is any capacitor that varies its capacitance by physically moving the location of its components (e.g., varying the distance between the plates, or the amount of plate surface area that overlaps). An MVC may be contrast with an electronically variable capacitor, such as those discussed herein that vary capacitance by switching in or out discrete capacitors.
Semiconductor Processing System and Matching Network
As is shown, the semiconductor device manufacturing system 40 utilizes an RF generator 30 (sometimes referred to as an RF source). The system 40 includes the RF generator 30 and a semiconductor processing tool 42. The semiconductor processing tool 42 includes a matching network 44 and a plasma chamber 20. In other embodiments, the generator 30 or other power source can form part of the semiconductor processing tool.
The semiconductor device can be a microprocessor, a memory chip, or other type of integrated circuit or device. A substrate 24 can be placed in the plasma chamber 20, where the plasma chamber 20 is configured to deposit a material layer onto the substrate 24 or etch a material layer from the substrate 24. Plasma processing involves energizing a gas mixture by imparting energy to the gas molecules by introducing RF energy into the gas mixture. This gas mixture is typically contained in a vacuum chamber (the plasma chamber 20), and the RF energy is typically introduced into the plasma chamber 20 through electrodes 22, 26. In processes that are well known in the art, the first and second electrodes 22, 26, in conjunction with appropriate control systems (not shown) and the plasma in the plasma chamber, enable one or both of deposition of materials onto a substrate 24 and etching of materials from the substrate 24.
In a typical plasma process, the RF generator 30 generates power at a radio frequency (RF) which is typically within the range of 3 kHz and 300 GHz—and this power is transmitted through RF cables and networks to the plasma chamber 20. In order to provide efficient transfer of power from the RF generator 30 to the plasma chamber 20, an intermediary circuit is used to match the fixed impedance of the RF generator 30 with the variable impedance of the plasma chamber 20. Such an intermediary circuit is commonly referred to as an RF impedance matching network, or more simply as an RF matching network. The purpose of the RF matching network 44 is to transform the variable plasma impedance to a value that more closely matches the fixed impedance of the RF generator 30. Commonly owned U.S. patent application Ser. No. 14/669,568, the disclosure of which is incorporated herein by reference in its entirety, provides an example of such a matching network.
As shown in
As discussed above, the RF impedance matching network 44 serves to help maximize the amount of RF power transferred from the RF source 30 to the plasma chamber 20 by matching the impedance at the RF input 101 to the fixed impedance of the RF source 30. The matching network 44 can consist of a single module within a single housing designed for electrical connection to the RF source 30 and plasma chamber 20. In other embodiments, the components of the matching network 44 can be located in different housings, some components can be outside of the housing, and/or some components can share a housing with a component outside the matching network.
As is known in the art, the plasma within a plasma chamber 20 typically undergoes certain fluctuations outside of operational control so that the impedance presented by the plasma chamber 20 is a variable impedance. Since the variable impedance of the plasma chamber 20 cannot be fully controlled, and an impedance matching network may be used to create an impedance match between the plasma chamber 20 and the RF source 30. Moreover, the impedance of the RF source 30 may be fixed at a set value by the design of the particular RF source 30. Although the fixed impedance of an RF source 30 may undergo minor fluctuations during use, due to, for example, temperature or other environmental variations, the impedance of the RF source 30 is still considered a fixed impedance for purposes of impedance matching because the fluctuations do not significantly vary the fixed impedance from the originally set impedance value. Other types of RF sources 30 may be designed so that the impedance of the RF source 30 may be set at the time of, or during, use. The impedance of such types of RF sources 30 is still considered fixed because it may be controlled by a user (or at least controlled by a programmable controller) and the set value of the impedance may be known at any time during operation, thus making the set value effectively a fixed impedance.
The RF source 30 may be an RF generator of a type that is well-known in the art, and generates an RF signal at an appropriate frequency and power for the process performed within the plasma chamber 20. The RF source 30 may be electrically connected to the RF input 101 of the RF impedance matching network 44 using a coaxial cable, which for impedance matching purposes would have the same fixed impedance as the RF source 30.
In the exemplified embodiment, the RF impedance matching network 44 includes a series VVC 34, a shunt EVC 36, and a series inductor 48 to form an ‘L’ type matching network. The shunt variable capacitor 36 is shown shunting to a reference potential, in this case ground, between the series VVC 34 and the series inductor 48, and one of skill in the art will recognize that the RF impedance matching network 44 may be configured with the shunt EVC 36 shunting to a reference potential at the RF input 101 or at the RF output 102. It is further noted that while the shunt variable capacitor 36 is an EVC in this embodiment, in other embodiments it may be another type of capacitor capable of varying its capacitance.
While the exemplified matching network 44 is in an L configuration, the matching network may be configured in other matching network configurations, such as a ‘T’ type configuration or a ‘Π’ or ‘pi’ type configuration. In certain embodiments, the variable capacitors and the switching circuit described below may be included in any configuration appropriate for an RF impedance matching network.
Each of the variable capacitors 34, 36 (as well as RF source 30) are connected to a control circuit 38, which is configured with an appropriate processor and/or signal generating circuitry to provide a signal for controlling the capacitors 34, 36. One or more power supplies (not shown) may be connected to components of the matching network 44 to provide operational power. It is understood that one or more of the variable capacitors may be operable coupled to choke, filter, and/or driver circuits for carrying out the functions described herein.
In the exemplified embodiment, the control circuit 38 includes a processor. The processor may be any type of properly programmed processing device, such as a computer or microprocessor, configured for executing computer program instructions (e.g., code). The processor may be embodied in computer and/or server hardware of any suitable type (e.g., desktop, laptop, notebook, tablets, cellular phones, etc.) and may include all the usual ancillary components necessary to form a functional data processing device including without limitation a bus, software and data storage such as volatile and non-volatile memory, input/output devices, graphical user interfaces (GUIs), removable data storage, and wired and/or wireless communication interface devices including Wi-Fi, Bluetooth, LAN, etc. The processor of the exemplified embodiment is configured with specific algorithms to enable matching network to perform the functions described herein.
The control circuit 38 is the brains of the matching network 44, as it receives multiple inputs, from sources such as the RF input sensor 32 and the variable capacitors 34, 36, and the RF source 30, and delivers commands to the variable capacitors 34, 36 (and RF source in the case of frequency tuning) to create the impedance match. The control circuit 38 is of the type of control circuit that is commonly used in semiconductor fabrication processes, and therefore known to those of skill in the art. Any differences in the control circuit 38, as compared to control circuits of the prior art, arise in programming differences, which, along with the matching network components and architecture, enable increased speed and reliability.
To compensate for these minor variations, since the matching network lacks future knowledge of the process to know where the VVC needs to be set, the exemplified control unit 38 is programmed to set the VVC to some predetermined position, and then change the VVC's capacitance to achieve a matched condition. It might also be necessary to have a narrow tuning range to have the efficiency increased. In this case, some variation would need to take place to match. In either case, the matching network needs to know how to accomplish this.
Process Variation Adjustment
According to a first method, the matching network takes fixed steps towards a VVC capacitance sufficient to cause the RF source frequency to be within a desired range. In one embodiment, according to a first process, tuning tables of a typical variable frequency EVC match are created, where an S-Map (an S-parameter map) is taken (other parameter matrices may alternatively be utilized). Each shunt capacitor position is swept with frequency. These datapoints are then used to generate tables where the unit can calculate the load impedance and then look up the best matching position for that impedance according to the first process. If the table finds that the calculated frequency is at the minimum or the maximum frequency for the RF source, the control unit will, according to a second process, reduce or increase the VVC capacitance until the input of the matching network is matched within the frequency range. The step size for the VVC can be as large or small as desired by the user. Since the best match frequency is at or outside of the edge of the S-Map, the units will always know which direction to adjust the VVC.
In this embodiment, while the VVC is being adjusted, the generator's frequency will be changed accordingly to either the minimum or maximum frequency, pursuant to the first process, to reduce the amount of reflected power, until the control unit determines that the best match frequency is within its frequency tuning range. Small steps would allow the VVC to adjust, and not overstep the needed change in capacitance. If a larger step is used, it is possible that the VVC would need to be readjusted if the process requires the control unit to go near or beyond the opposite extreme during a later step in the process. It would not be a problem to match once the load impedance is in range. The RF source would be able to continuously shift frequency while the VVC is still adjusting, once the load impedance is within the frequency tuning range. A larger step may allow the VVC to only be adjusted once, allowing the best match to be achieved faster, because the VVC is not adjusted multiple times. This balancing act would need to be optimized on the tool and could be different depending on the process.
According to another method, the matching network takes fixed step and knows that the load is outside of the tuning range. For example, S-parameters may be taken with two additional points, one below the minimum operating frequency, and one above the maximum operating frequency, for each shunt capacitor position. When the control unit determines the best matching frequency is one of these outside frequencies, it will know to change the capacitance, instead of automatically changing when it might not be necessary. These outside points could be spaced such that when they are found to be the best match, the generator will see some minimum VSWR. This would reduce the tendency for capacitor adjustments. This would have to be already calculated in the table.
According to another method, the steps are a ratio of the frequency outside of the desire range. One of the drawbacks of the previous two methods to compensate for when the matching network is unable to achieve a perfect match is that it may take more than one step to put the matching network in the proper range. It may take several steps depending on what the step size is and how much compensation is needed. This could cause a significant delay in matching.
An improvement to these methods would be to measure additional S-parameters outside the frequency range of the matching network and take a guess as to how much the control unit needs to adjust the capacitor. This would allow the unit to know about how much to adjust the capacitor. A simple ratio could be used, such as 0.1 pF/1 MHz, for example. More complicated equations could be used based on what is found to work with the matching network. Based on what the user finds, a curve could be fit to optimize the tuning speed. This could be a more general equation that would work well for a whole product line, or it could be calibrated with each unit.
Fully Calculated Compensation
According to another method, an equation is used to determine step size. Specifically, the range is adjusted to calculate exactly the change in capacitance needed. This is more complicated. It requires the unit to know its VVC's capacitance at all possible setpoints and the inductance. The unit would then know, based on what the calculated best match frequency is outside of the tuning range, what the new capacitance should be to achieve a perfect match. If the capacitance is known, the inductance can be calculated by finding the resonant frequency, where the inductor and capacitor cancel out each other's impedances, leaving only the resistive part. This can be found in Equation 1 (Inductance Calculation), where L is the inductance, C is the capacitance, and f is the resonant frequency.
Multiple capacitance values can be used to vary the resonant frequency. These different calculated inductance values can then be averaged together to improve the accuracy.
Next, the relationship between the calculated frequency outside the range and the new capacitance may be found. Equation 2 (Impedance Calculation) shows how to calculate the impedance of the circuit, X. It is assumed that the resistances are negligible, and do not have a large effect on the calculation.
The new impedance (X′) can be calculated by substituting the next frequency step (F′) that is outside of the range and the present capacitance value C. This would be equivalent to adjusting the capacitor to some new value (C′) when the frequency (F″) is set to either the minimum or maximum, whichever is closest, to generate the same impedance. This can be found in Equation 3 (Impedance Equivalence).
C′ can then be solved. This is shown in Equation 4 (New Capacitance). Note that another frequency within the tuning range may be used to add some buffer, which may prevent the match from needing to retune the capacitor. An example would be the maximum frequency minus one percent of the tuning range, or the minimum plus one percent.
Another example of what the calculated method could be used for is centering the frequency range. While a process is running, it may be desirable to adjust the frequency to its midpoint, giving the maximum margin for frequency tuning. Some examples may be used to maximize the matching speed when there is some variability in the process, between process steps, when changes are unknown, or if certain process steps cause transients. The calculation for this is like before, as seen in Equation 5 (Impedance Equivalence) and Equation 6 (New Capacitance) below. Here, F and C are the present frequency and capacitance and Fmid and C′ are the center frequency and the new capacitance value. While the VVC is changing its position, the control unit can adjust the frequency as needed to allow for minimum reflected power throughout the transition.
The foregoing calculated method could also be used to implement a matching window. While the process is running, frequency would have a keep out area, for example less than 25% and greater than 75% of the frequency tuning range. This could also be hysteretic where the tuning will adjust the VVC to some position to have the frequency closer to the center. For example, if the frequency hits its lower or upper 25%, the VVC will be adjusted so that the calculated frequency is tuned to 30% or 70%.
Calculating the Load Impedance
It may be desired, for various reasons, to calculate the load impedance. There are a few different ways to accomplish this, which includes different setups and desired accuracies. Referring back to
For the first case, the series VVC 34 capacitance is known. After the load impedance is calculated, which includes series VVC 34, the capacitor's impedance can be simply subtracted from the load. The only information needed is the capacitance of series VVC 34, and the known frequency. This isn't the most accurate method, as it does not include parasitics and other elements that may be in the circuit between series inductor 48 and the output 102.
The next method, which is more accurate, is to take S-parameters between the node between series VVC 34 and series inductor 48, and the output 102. These parameters can be used to de-embed series VVC 34 from the load 20. This would require more complex calculations than the previous method. It also requires a table and additional memory of series VVC 34 position versus frequency, for all the desired capacitances and frequencies.
The second case is more difficult to conceptualize. The S-Map Data is collected with series VVC 34, and the change in series VVC 34 will appear to be a change in the load impedance. This change will not impact the auto matching, but it will create an erroneous load impedance calculation.
The first method will use the change in capacitance of series VVC 34 to determine the actual load impedance. To do this, the series VVC 34 capacitance must be known when the S-Map data is collected. This will be used as a fixed capacitance. Changes to series VVC 34 are equivalent to having a variable capacitor (CP) in parallel with series VVC 34. Equivalently, a variable capacitor (CS) can be placed in series with series VVC 34.
The above relationship between is expressed in Equation 7 (Parallel to Series Capacitors Equivalence) below. Solving for CS can be found in Equation 8 (Calculate for CS). The impedance of CS can now be calculated and subtracted from the load impedance.
S-parameter can be used to give a more accurate load calculation using the second case. To do this, series VVC 34 must be known when the S-Map data is collected. Next, S-parameters of series VVC 34 need to be taken, from the node between series VVC 34 and series inductor 48, and the output 102. During run time, these parameters will be used to de-embed the original series VVC 34 setting, which was the setting during S-Map, from the output S-Map data for the current match settings, and embed the current series VVC 34 setting. The match would then calculate the actual load impedance.
Another circuit that would work well with the previous methods would be a parallel inductor and capacitor. The S-Map data could be taken in the same ways, and the adjustments be could preformed in any of the previously mentioned tuning methods. The only changes would be the equations in the fully calculated compensation.
For the fully calculated version of the parallel circuit, Equation 1 (Inductance Calculation) could be used. Equation 10 (Parallel Impedance Calculation) below shows how the impedance is calculated for a parallel LC network. The equivalence of changing the frequency versus changing capacitance is found Equation 11 (Parallel Impedance Equivalence), where the variables are similar to Equation 3 (Impedance Equivalence). Finally, the new capacitance is found in Equation 12 (New Parallel Capacitance).
There are many other matching network circuits that could be designed to utilize the methods discussed above.
The method 81 includes a first process 81A and a second process 81B. According to the first process 81A, a sensor configured measures a parameter related to the plasma chamber (operation 82). The parameter may be, for example, a voltage, a current, and/or a phase at the RF input of the matching network, or a forward and/or reflected power at the RF input of the matching network. The invention, however, is not so limited, as the parameter may be any parameter related to the plasma chamber.
Next, the control circuit determines a parameter-based value based on the measured parameter (operation 83). In one embodiment, the parameter-based value is a load impedance value, the measured parameter value is used to determine an input impedance value, and the input impedance value is used to determine the load impedance value. Further, the load impedance value may be determined by entering an impedance value for the input of the matching network into a first parameter matrix, such as an S-Map. The invention, however, is not so limited, as the parameter-based value may be any value based on the measured parameter (including the measured parameter itself), and a parameter matrix need not be utilized. It is understood that, while S-parameters and S-Maps are discussed in certain embodiments herein, such may be replaced with other types of parameter matrices. It is further noted that any of the parameter matrices discussed herein may form part of a parameter look-up table.
Next, the control circuit, using the parameter-based value (e.g., load impedance), determines a second variable capacitor configuration and an RF source frequency for reducing a reflected power, and then sends a signal to alter the second variable capacitor to the second variable capacitor configuration and the RF source to the RF source frequency (operation 84). In one embodiment, the second variable capacitor configuration and the RF source frequency for reducing the reflected power are determined by inputting the parameter-based value into a second parameter matrix, though the invention is not so limited.
As for the second process 81B, first, the RF source frequency is determined (operation 86A). Next, there is a determination whether the alteration of the RF source frequency has caused the RF source frequency to be outside, at a minimum, or at a maximum of a predetermined frequency range (operation 86). If so, the control circuit determines a new MVC configuration to cause the RF source frequency, according to the first process 81A, to be altered to be within or closer to the predetermined frequency range, and sends a signal to alter the MVC to the new MVC configuration (operation 87). The determination of the new MVC configuration is based on the RF source frequency and the predetermined frequency range. The determination of the new MVC configuration may also be based on the second variable capacitor configuration. In one embodiment, the new MVC configuration is determined using a parameter matrix with a look-up table, though the invention is not so limited.
In certain embodiments, the predetermined frequency range comprises a plurality of frequency values, the alteration of the RF source frequency has caused the RF source frequency to be at the minimum or at the maximum of the plurality of frequency values, and the new MVC configuration causes the RF source frequency, according to the first process, to be altered to be within the predetermined frequency range such that the RF source frequency is no longer at the minimum or at the maximum of the plurality of frequency values. In other embodiments, the predetermined frequency range is a single frequency value (e.g., 13.56 MHz), the alteration of the RF source frequency has caused the RF source frequency to not be at the single frequency value, and the new MVC configuration causes the RF source frequency, according to the first process, to be at or closer to the single frequency value.
As shown in
As shown above, there are multiple ways to adjust the tuning range of a matching network utilizing and a variable capacitor (such as an EVC), frequency tuning, and an MVC (such as a VVC), which may further include use of a parameter matrix (such as an S-Map). The matching network can quickly calculate the need for a change and then adjust of the tuning range with the VVC if necessary. This achieves very fast tuning with a minimal frequency tuning range to reduce inefficiencies.
In certain embodiments, the matching networks discussed above may form part of a semiconductor processing tool (such as tool 42) that comprises a matching network and a plasma chamber. Further, the methods of impedance matching discussed above may form part of a method of manufacturing a semiconductor.
As discussed above with regard to frequency tuning during impedance matching using VVC adjustment, matching networks that have control of the frequency of the RF source can provide several advantages, including cost advantages. But there are potential disadvantages, such as a reduction in efficiency and/or a narrow tuning range, which are discussed above with respect to
Adding a variable capacitor requires that the tuning range is only as wide as the process that covers the largest area, as shown in plot 18 of
Another way to further increase the efficiency without sacrificing tuning range or increasing cost is to measure the S-Map with quantized values of the variable capacitor, adding an extra dimension to the S-Map. As used herein, the term “quantized” refers to using multiple discrete values for the variable capacitor in question. Using quantized or discrete values adds an additional dimension to the S-Map because, instead of ignoring the fact that the variable capacitor can change and measuring S-parameters only for one value of the variable capacitor, the system can measure multiple S-Maps, each for a different value of the variable capacitor. According to this approach, each variable capacitor would be set to some values when measured, and the variable capacitor would not be able to change between values during operation. This would allow for a decrease in the series inductor (such as series inductor 48), which would increase the efficiency. Such an approach will increase the size of the tuning tables derived from the S-Map, because an additional dimension is added, requiring more memory (e.g., by a factor of the number of quantized steps). The required memory may be reduced, however, by reducing the number of frequency points taken in the S-Map. Since the area for each capacitance has been reduced, the same resolution can be found. The number of points may be reduced by the same factor that the number of points were increased to achieve the same resolution, but it is recommended that there is some overlap to avoid gaps in the S-Map.
As discussed above, the variable capacitor may be a VVC or an EVC. An EVC has an advantage over a VVC in that an EVC can adjust itself extremely quickly, potentially thousands of times faster than a VVC. This is because VVCs are a mechanical device that need to spin a shaft to vary the capacitance, while EVCs use solid state switches. These switches can adjust the full range of capacitance in one control cycle. Typically, VVCs are set to a predetermined capacitance to work effectively. Otherwise, it could take a significant amount of time to match. An EVC typically uses many switches, each corresponding with an individual capacitor, to produce a wide range of values without creating large gaps in capacitance. These large gaps would create quantization errors, in the sense that the ideal capacitor value is not achievable, as it is with a traditional VVC with potentially unlimited resolution. Thus, many switching elements, drivers, and capacitors are required to minimize this error, which require additional cost.
With the new multi-dimensional S-Map, however, the number of switched capacitors for an EVC can be reduced significantly. Only coarse step capacitors would be needed, because fine tuning can be accomplished using frequency tuning. For example, if series variable capacitor 34 was an EVC, only 3 coarse capacitors could be used to switch in and out (e.g., 3 coarse capacitors of 7.5 pF each). This could reduce the cost of the EVC enough that it becomes a cost advantage over a VVC. Alternatively, if series variable capacitor 34 was a VVC, the stepping size could be every 7.5 pF (providing, e.g., 90 pF, 97.5 pF, and 105 pF), instead of a more typical step size of every 0.1 pF. The large step size can cause large gaps between values in the S-Map, but these large gaps can be filled in by the frequency adjustment.
The variable capacitor that is added to the frequency tuning matching network allows for a wider tuning range to be attained without increasing loss. The variable capacitor can be added to the S-Map calibration at various quantized capacitance values, enabling the matching network to quickly tune the circuit by varying both frequency and capacitance. This method could also be used to reduce the cost of an EVC matching network by reducing the number of switched capacitors needed.
There are many circuits that could be designed to utilize this method. For example, in addition to the matching network 44 shown in
Next, the parameter-based value (e.g., load impedance) is input into a match configuration look-up table to determine a match configuration for reducing a reflected power (operation 113). The match configuration includes a first variable capacitor configuration, a second variable capacitor configuration, and an RF source frequency for reducing a reflected power. Next, the first variable capacitor is altered to the first variable capacitor configuration, the second variable capacitor is altered to the second variable capacitor configuration, and the RF source is altered to the RF source frequency to cause a reduction of the reflected power.
In certain embodiments, the matching networks discussed above may form part of a semiconductor processing tool (such as tool 42 of
Parameter Matrices
The exemplified impedance matching network may be characterized by one of several types of parameter matrices known to those of skill in the art. An S-parameter matrix and a Z-parameter matrix are two examples of such parameter matrices. Other examples include, but are not limited to, a Y-parameter matrix, a G-parameter matrix, an H-parameter matrix, a T-parameter matrix, and an ABCD-parameter matrix. Those of skill in the art will recognize also that these various parameter matrices may be mathematically converted from one to the other for an electrical circuit such as a matching network.
A parameter look-up table includes a plurality of parameter matrices, with each parameter matrix being associated with a particular configuration of a variable capacitor and/or RF source. In embodiments in which the parameter look-up table includes multiple types of parameter matrices, the different types of parameter matrices may be associated within the parameter look-up table in such a way so as to eliminate the need for mathematical conversions between the different types of parameter matrices. For example, the T-parameter matrix may be included as part of the parameter look-up table, with each T-parameter matrix associated with the associated S-parameter matrix that would result from conversion between the two matrices.
The match configuration look-up table is a table of match configurations for the variable capacitors and RF source, and it may include each possible array configuration of the variable capacitors, and each possible frequency configuration for the RF source. As an alternative to using a match configuration look-up table, the actual capacitance values or switch configurations for each variable capacitor (and the actual RF frequency for the RF source) may be calculated during the process, such real-time calculations of the capacitance values are inherently slower than looking up the match configurations in the match configuration look-up table.
The altering of the variable capacitors and RF source may include a control circuit sending one or more control signals to cause the change to the match configuration. For example, a control signal may be sent to a driver circuit to control a capacitor configuration. When the variable capacitors and the RF source are switched to the match configurations, the input impedance may match the fixed RF source impedance (e.g., 50 Ohms), thus resulting in an impedance match. If, due to fluctuations in the plasma impedance, a sufficient impedance match does not result, the process may be repeated one or more times to achieve an impedance match, or at least a substantial impedance match.
The look-up tables used in the process described above are compiled in advance of the RF matching network being used in conjunction with the plasma chamber. In one embodiment, in creating the look-up tables, the matching network is tested to determine at least one parameter matrix of each type and the load impedance associated with each match configuration of the variable capacitors and RF source prior to use with a plasma chamber. The parameter matrices resulting from the testing are compiled into the parameter look-up table so that at least one parameter matrix of each type is associated with a respective match configuration of the variable capacitors and RF source. Similarly, the load impedances are compiled into the match configuration look-up table so that each parameter matrix is associated with a respective match configuration of the variable capacitors and RF source. The pre-compiled look-up tables may take into consideration the fixed RF source impedance (e.g., 50 Ohms), the power output of the RF source, and the one or more operational frequencies of the RF source, among other factors that are relevant to the operation of the RF matching network. Each look-up table may therefore have tens of thousands of entries, or more, to account for all the possible configurations of the variable capacitors and RF source. The number of possible configurations is primarily determined by how many discrete capacitors make up each of the variable capacitors and the number of possible frequency values provided by the RF source. In compiling the look-up tables, consideration may be given to possible safety limitations, such as maximum allowed voltages and currents at critical locations inside the matching network, and this may serve to exclude entries in one or more of the look-up tables for certain configurations of the variable capacitors or RF source.
As is known in the art, the S-parameter matrix is composed of components called scatter parameters, or S-parameters for short. An S-parameter matrix for the impedance matching circuit has four S-parameters, namely S11, S12, S21, and S22, each of which represents a ratio of voltages at the RF input and the RF output. All four of the S-parameters for the impedance matching circuit are determined and/or calculated in advance, so that the full S-parameter matrix is known. The parameters of the other types of parameter matrices may be similarly determined and/or calculated in advance and incorporated into the parameter matrix. For example, a Z-parameter matrix for the impedance matching circuit has four Z-parameters, namely Z11, Z12, Z21, and Z22.
By compiling the parameter look-up table in this manner, the entire time cost of certain calculations occurs during the testing phase for the RF matching network, and not during actual use of the RF matching network with a plasma chamber. Moreover, because locating a value in a look-up table can take less time than calculating that same value in real time, using the look-up table can aid in reducing the overall time needed to achieve an impedance match. In a plasma deposition or etching process which includes potentially hundreds or thousands of impedance matching adjustments throughout the process, this time savings can help add directly to cost savings for the overall fabrication process.
Filters for Switching Circuit
As EVC matching networks become more prevalent, end users will seek to push the limits of the technology. The matching speed of EVC matching networks far surpasses that of the traditional VVC matches, enabling applications previously unavailable. An example would be level-to-level pulse matching, where power is pulsed but the state is either high or low, rather than ON or OFF. The RF source generates a multi-level pulse signal such that the RF signal has cyclically recurring pulse intervals with differing amplitude levels. In some cases, the change in the power setpoint amplitude level can be very frequent and of the order of a few tens of hundreds of microseconds. An example of a matching network using level-to-level pulsing is described in co-owned U.S. patent application Ser. No. 16/592,453, which is incorporated herein by reference in its entirety.
Such a level-to-level pulsing application would require the matching network to be in a state of continuous matching. This is because the match is required to match for a short duration for either the high or low pulse. Then it will be required to match again when the next pulse state is initiated. Depending on the period and duty cycle, this may require the match to switch continuously.
As discussed above, EVCs comprise several discrete capacitors. In EVC-based matching networks, PIN (or NIP) diodes may be used as the RF switching device for switching in and out the discrete capacitors of the one or more EVCs in the matching network. In the matching network 90A of
Referring again to
But this choke design also has drawbacks. The main problem is that it affects the overall switching speed, where large voltages can be generated across the inductor. According Equation 1 below, the voltage across the inductor 64A is equal to the inductance times the rate of change in current. The current comes from the Bias 68A and therefore is fixed. If the inductance of the choke needs to be large, then the Bias being applied must gradually increase and decrease.
As an example, if the inductance is 60 uH, the bias current is 0.5 Amps, and the switching speed is 10 ns, the voltage generated across the inductor would be 3000 Volts. This may damage the filter 62A, driver circuit 66A, PIN diode 63 and/or discrete capacitor 78. This voltage also oscillates and will cause the PIN diode 63 to turn ON and OFF rapidly. If the switching speed is changed to 50 us, this voltage drops to only 0.6 Volts. To drop the inductance by the same magnitude would also have a similar effect, but the isolation required prevents this.
Filtering capacitor 62A is a lowpass filter that helps block the RF from the driver circuit. This capacitance is typically a large value (e.g., 3000 pF for a 13.56 MHz operating frequency) to give a low impedance to ground for the RF signal. This large capacitance, however, adds stress to the driver circuit 66A by creating additional power dissipation. The power dissipation is difficult to calculate because of the complexity of the circuit, especially with the slow, nonlinear transition times of the driver circuit. We know the energy stored in the capacitor and the minimum amount of time that it takes to fully charge and discharge, however, so the power dissipated per charge and discharge can be roughly calculated for the full system.
Equation 2 shows how to calculate the stored energy in the filtering capacitor 62A. Equation 3 shows how to calculate power dissipation from the energy dissipated over time. If Equation 2 is substituted into Equation 3 then the total power dissipated in the system from the capacitor charge/discharge can now be calculated as seen in Equation 4.
Using Equation 4, the total power dissipation per transition can be calculated. To continue with the previous example, the switching speed is 50 us. If the HVDC is set to 1650V and the capacitance is set to 3400 pF, then the total power dissipated in the system is about 92.6 W. This dissipation includes losses in the driver circuit switch, the power supply, the bias resistor, and anything else in series with the capacitor while it either charges or discharges.
The power dissipation will generate a considerable amount of heat. The heat will degrade components and reduce their lifetime. This will require the match to reduce the number of transitions in a given amount of time.
Another issue is that the HVDC supply needs to be able to supply current for multiple switches. In the previous example, the HVDC 67A would be applied to the PIN diode 63 every other OFF transition. Thus, the supply would need to power 46.3 W. For a worst-case scenario, this would then need to be multiplied by the number of switches. If the matching network has 50 PIN diodes, for example, the HVDC supply would need to be rated for over 2300 Watts. This calculated power requirement is a worst-case example, as the probability of the switches needing to transition from all OFF to all ON continuously is very low.
A way to help with the required power is to place a very large capacitance on the HVDC supply. This will work for non-level-to-level processes or very short level-to-level processes. Using the worst-case scenario, assuming the HVDC supply was able to supply 23 Watts, or half of a switch, the unit would need to switch in and out 3400 pF multiplied by 49.5 switches per every OFF transition. If a 10 uF capacitor is used to store charge, with 100 us between off transitions, and a minimum voltage of 1600 VDC was required, the switches would be able to have two OFF transitions. This would give a total time of 200 us before stopping. For the best-case scenario, the unit would only change a single switch or have 0.5 switches per transition. In this case, the unit would be able to make 182 off transitions, or run continuously for 18.2 ms, before the HVDC drops below 1600 VDC.
The plots of
The HVDC supply capacitor can be increased, but at a certain point the size of the capacitor, or group of smaller capacitors, will become very large and would have a significant cost. Another option is to simply reduce the capacitance of the filter capacitors. This option is not ideal, however, because the filter response is critical to the operation of the driver circuit. Without it, EMI would cause the driver circuit to not operate properly.
Band-Stop Filter for Switching Circuit
In the exemplified embodiment, for a 13.56 MHz matching network, the filtering capacitor 62C has a capacitance of 300 pF, and the inductor 65C has an inductance of 470 nH. In other embodiments, these values may differ. For example, in certain embodiments, the filtering capacitor may have a capacitance of 110 pF with an inductor of 150 nH to filter 40.68 MHz. The exemplified band-stop filter 67C is parallel to the PIN diode 63, and its inductor 65C is coupled between the filtering capacitor 62C and the common ground 43. In other embodiments, the filtering capacitor may be coupled between the inductor and the common ground. In this embodiment, the biasing current supply 68C and the band stop filter circuit 67C are coupled to the common ground 43. The exemplified PIN diode 63 is also coupled to the common ground 43, though the invention is not so limited.
Similar to
As discussed above, a reduction in the capacitance of the filtering capacitors 62C for each switching circuit 61C would reduce the power dissipation. Using the previous example but allowing the capacitance to drop by a factor of 10, the power supply would now be able to handle switching five capacitors continuously, and thus the best-case scenario is switching five capacitors or less. If a sixth capacitor were to be switched continuously, the time to drop the HVDC to below 1600 VDC would be 93.3 ms. If the original capacitance was used, the time to drop the HVDC with six switches would be 1.7 ms. The worst-case scenario with the new circuit would drop the HVDC in 2.1 ms.
A potential disadvantage of using band-stop filters would be that the filters may need to be tuned or tightly controlled. The bandwidth where the minimum attenuation is met would be decreased, and thus the frequency range would be reduced. An example simulation of the two filter responses can be seen in
To increase the number of switches that can switch continuously, following the previous example, the total capacitance must be 1690 pF. That would be divided by the number of PIN diode switches. For a fifty-switch system, the capacitance would need to drop to 33.8 pF per switch. This may cause an undesirably narrow bandwidth, and therefore some balance must be made between the minimum number of switches that must continuously switch, and the attenuation bandwidth of the filter.
The above-described filter topologies may be used as part of a method for providing an impedance match, where an impedance is matched by at least one of the switching circuits of the EVC switching in or out its corresponding discrete capacitor to alter a total capacitance of the EVC. Further, this method of matching an impedance can form part of a method for fabricating a semiconductor, such as those methods for fabricating a semiconductor discussed above.
In view of the foregoing, changing the filter topology from a lowpass filter topology to a band-stop filter topology has multiple advantages because of the drop in the filtering capacitor. First, the band-stop filter topology reduces the power dissipation in the switching devices. This will reduce their temperature rise, allowing them to switch more often and increase their lifetime. Second, because of the lower power consumption, the power supply will allow for more switches to be switched simultaneously and more often. This design can also be modified to further increase the number of transitions. If the capacitance is dropped enough, it will no longer be the limiting factor in switching time. At that point, the RF choke will need to be reduced, and a discrete capacitor will need to be added to shift its resonant frequency. These changes will increase the switching speed, allowing for faster pulsing and matching.
Adding an inductor to the filtering capacitor to create a band-stop filter for the impedance matching network provides a new and effective means of increasing the effective capacitance of the filter without adding to the charging and discharging capability of the filter.
FET-Based Switching Circuit with Voltage Bias to Reduce Parasitic Capacitance and Capacitance Variability
Typically, solid state matching networks use one of two different switching elements. The first is a PIN diode, which is controlled by either a forward bias current, to turn the switch on, or a reverse bias voltage, which turns the switch off. The other is a transistor, such as a Field Effect Transistor (FET). A FET is driven by low and high voltages to turn the devices off and on.
The main advantage of using FETs over PIN diodes is that they have a faster switching speed. This allows the matching network to match more quickly and have lower switching losses. FETs can also have a lower equivalent series resistance (ESR) per switch. There is also a larger selection of parts and more availability. The number of components needed is also reduced, as the transistor circuit is what essentially drives the PIN diode.
One disadvantage of using FETs is that the parasitic capacitance is higher than when using PIN diodes. This makes the FET's impedance at RF frequencies have less of a dynamic response when switched between the on and off states. Another disadvantage is related to the FETs' inherent design. A FET has a body diode across its drain and source. This body diode leads one to use two FETs in series with either a common source or a common drain, as shown in
It is noted that for the sake of clarity and convenience in describing similar components or features, the same or similar reference numbers may be used herein across different embodiments or figures. This is not to imply that the components or features identified by a particular reference number are identical across each embodiment or figure, but only to suggest that the components or features are similar in general function or identity.
A simulation was performed on the two equivalent circuits 32A-E, 32B-E of
Unlike PIN diodes, the bias in the off state varies the capacitance over a wide range. The capacitive variation from the change in amplitude in existing designs will cause there to be a large difference between the S-parameter map (S-Map), which is unmatched at low power, meaning the voltage on the switches is extremely low, as opposed to when the unit is under power and matched, generating voltages that can be near or at the limit.
Design 1: Voltage Bias Across Drain-to-Souce of FET
In a first design embodiment described below, a switching circuit will have an applied voltage bias across the drain-to-source of the switching FET to reduce capacitance variability. The bias is applied when the FET is in the off condition, like the reverse bias to turn off a PIN diode. The description below will then show different topologies to allow the switch to operate at higher voltages.
As shown in
Several advantages are provided by Design 1. If a DC voltage was applied across the switch at half the rated voltage, or half the maximum allowable RF peak voltage, the variation in capacitance due to changes in amplitude would be mitigated. This means that the impedance of the switching elements would be constant when taking the S-Map measurement and when applying the maximum allowable peak voltage to the device. The stability in the FETs' impedance is critical for any matching network, including those that do not utilize an S-Map. Without this, as the matching network approaches a matched condition, there is less reflected power. Therefore, the voltage on the capacitor array will increase. Thus, as a capacitor is switched in to increase the total capacitance, the voltage would be increased, which will reduce the capacitance of all the switches in the off position. This could cause an oscillation, as the capacitance would fluctuate during tuning. The other condition is where the capacitance is switched out of circuit to reduce the total capacitance. The voltage would again increase dropping the capacitance further, creating an overcompensation. This would make it difficult to predict how much to reduce the capacitance by, requiring a smaller step than what is necessary.
Another benefit of Design 1 is to reduce the parasitic capacitance of each switching circuit, which will increase its impedance in the off state, making it look more like an ideal switch. This will have two effects. The first is that it will create the lowest possible starting capacitance, creating a broader tuning range. The second, from what has been observed in literature and from Applicant's testing, is that the efficiency tends to be higher with lower capacitance. This is from the higher impedance having a lower current, reducing FR losses.
This design would be simpler to implement than a switch with an amplitude variable inductor in parallel with each switching element. Now, a fixed inductor could be placed in parallel to further reduce the parasitic capacitance if needed. It will also have lower losses than using one or more diodes in series with the FET, creating a more efficient match.
The biggest challenge with this design is that the DC voltage takes away from the maximum RF peak. The maximum RF voltage that can be applied is less than or equal to the HVDC bias. Ideally, half the rated voltage would be applied to the FET as bias, which means the maximum RF peak is reduced to half of the rated voltage. Design 1 also requires high-voltage direct current (HVDC) to implement this design, increasing the probability of arcing and thus requiring certain precautions to be made.
In one embodiment, Design 1 utilizes a half H-bridge configuration. A bottom FET will be the RF switching element, and a top FET will be used to bias the bottom FET when it is in the off state. Note, this is just one implementation, as there are many ways to apply the bias voltage. These switch designs can also be in the series configuration as stated previously.
In certain embodiments, a blocking element is placed between the RF switches and the HVDC supply to block RF and pass HVDC. If not, the RF current may see a low impedance path to ground, giving an apparent impedance drop of the bottom FET in the off state. There are many ways of doing this, such as using chokes, diodes, or large-value resistors on each channel, or some combination of any of these per channel. As will be shown, a single element, or combination of elements, from the previously mentioned can be used for all channels.
If there is isolation between the top FET (the biasing FET) and bottom FET (the switching FET), there will not be any considerable RF current flowing through the top FET. This means that the top FET (biasing FET) only needs to have half the voltage rating of the bottom FET (switching FETE), HVDC with some margin. Thus, a much less expensive low current switching element could be used to apply the bias.
The advantage of using a blocking resistor as shown is that the circuit will have a very wide bandwidth. Thus, one switch design can be used for various frequencies or a multi-frequency unit. The difficulty with this design is it requires a resistor that is large enough to limit the current flowing from the RF to the HVDC supply, while also being small enough to charge the FETs' various capacitances in a short amount of time. When the switches are on, there will be a steady flow of current from the HVDC supply to ground. This implementation would have a slower switching response than other designs provided herein, but would be less expensive and easier to implement because of the simplicity of the design.
The higher RF isolation between channels 1228 will reduce the capacitance for each individual channel in the off case. This will drop the minimum capacitance. The low impedance for the HVDC 1212 will give the fastest possible switching. It's possible that the switching speed will cause some ringing on the inductor 1222. To reduce this, a resistor can be placed in series, slowing down the switching speed of the FETs. The resistor value can be much larger than in a typical PIN diode design, because there is no forward bias current.
Since all the channels 1228 are sharing one RF blocking element 1228, there is no isolation between channels. There are some things to consider because of this. The top biasing FET 1213 is the same as the switching FET 1214, because it will see the HVDC 1212 plus the full RF voltage. Further, the parasitic capacitance of all the channels 1228 will be in parallel. Thus, whether the other channels are on or off, there will be a FET's parasitic capacitance in parallel with the other channels that are off. This may reduce the range of the matching network.
It is possible to place multiple FETs in series to increase the maximum peak voltage. In the embodiments below, multiple gate drivers 1230, 1231 are be used. An input to the isolated driver was added to the left side of the driver to show how it is controlled by the half-bridge gate driver with programmable deadtime.
If there are two switching FETs 1214, the biasing FET 1213 would be chosen to have the same voltage rating. More FETs could also be stacked on the top if an even higher voltage is necessary or desired. If there are three or more switching FETs 1214, this would require more voltage on the biasing FET 1213. Since there is little to no RF current, the biasing FET 1213 could be a less expensive part with a lower voltage rating. With this scenario, it may be a cost advantage to have multiple biasing FETs 1213. A challenge of this setup is that the sharing of voltages may not be uniform across FETs. For example, the RF voltage, due to parasitic capacitance to ground 1243 from the layout, could have a larger drop across the first switching FET, than the subsequent switching FET(s) closer to ground 1243. The DC voltage may also take time to trickle down to the FETs closer to ground when the FETs are switched to the off state. These unequal sharing may cause damage to the FETs.
Design 2: Voltage Bias to Common Node of FETs
The Design 2 described below takes Design 1 with the opposing FETs and adds the bias voltage to the FETs' common node.
This circuit can be implemented in series mode as well.
Specifically,
Design 2 has several advantages. The applied bias will keep it from varying with amplitude and reduce the parasitic capacitance to a minimum. The HVDC bias can be set to be equal to the offset that is generated by the rectification of the body diode at the maximum RF peak voltage. Thus, unlike the Design 1, there is no reduction in the maximum RF voltage from the existing design.
The above-described switching circuit and EVC designs may be used as part of a method for providing an impedance match, where an impedance is matched by at least one of the switching circuits of the EVC switching in or out its corresponding discrete capacitor to alter a total capacitance of the EVC. Further, this method of matching an impedance can form part of a method for fabricating a semiconductor, such as those methods for fabricating a semiconductor discussed above. Further, the exemplified switching circuits and EVCs may form part of a matching network, which together with a plasma chamber may form part of a semiconductor processing tool, similar to those semiconductor processing tools discussed herein.
As used throughout, ranges are used as shorthand for describing each and every value that is within the range. Any value within the range can be selected as the terminus of the range. In addition, all references cited herein are hereby incorporated by referenced in their entireties. In the event of a conflict in a definition in the present disclosure and that of a cited reference, the present disclosure controls.
While the invention or inventions have been described with respect to specific examples, those skilled in the art will appreciate that there are numerous variations and permutations of the above described invention(s). It is to be understood that other embodiments may be utilized and structural and functional modifications may be made without departing from the scope of the present invention(s). Thus, the spirit and scope should be construed broadly as set forth in the appended claims.
Claims
1. An impedance matching network comprising:
- a radio frequency (RF) input operably coupled to an RF source providing an RF signal;
- an RF output operably coupled to a plasma chamber; and
- an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance of the EVC;
- wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor;
- wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a variable bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET; and
- wherein the bias voltage is varied based on a measured value of the RF signal to keep an offset at or slightly above a desired bias voltage.
2. The matching network of claim 1 wherein each switching circuit further comprises a blocking element positioned between the bias voltage source and the at least one switching FET, the blocking element blocking an RF current and passing the bias voltage.
3. The matching network of claim 2 wherein the blocking element comprises an inductor, a diode, or a resistor.
4. The matching network of claim 1 wherein the at least one switching FET comprises a first switching FET and a second switching FET coupled in series.
5. The matching network of claim 4 wherein either:
- a drain of the first switching FET and a drain of the second switching FET are coupled to a common node, the common node receiving the bias voltage; or
- a source of the first switching FET and a source of the second switching FET are coupled to a common node, the common node receiving the bias voltage.
6. The matching network of claim 1 further comprising a blocking inductor positioned between the bias voltage source and the switching circuits, the blocking inductor blocking an RF current from returning to the bias voltage source.
7. The matching network of claim 1 wherein each switching circuit further comprises at least one biasing FET positioned between the bias voltage source and the at least one switching FET, the biasing FET providing the voltage bias to the at least one switching FET.
8. A method of matching an impedance comprising:
- coupling an RF input of a matching network to an RF source providing an RF signal;
- coupling an RF output of the matching network to a plasma chamber, wherein the matching network comprises an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit configured to switch in and out the discrete capacitor, wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor;
- for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receiving a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET, the bias voltage being varied based on a measured value of the RF signal to keep an offset at or slightly above a desired bias voltage; and
- matching an impedance by at least one of the switching circuits of the EVC switching in or out its corresponding discrete capacitor to alter a total capacitance of the EVC.
9. The method of claim 8 wherein each switching circuit further comprises a blocking element positioned between the bias voltage source and the at least one switching FET, the blocking element blocking an RF current and passing the bias voltage.
10. The method of claim 9 wherein the blocking element comprises an inductor, a diode, or a resistor.
11. The method of claim 8 wherein the at least one switching FET comprises a first switching FET and a second switching FET coupled in series.
12. The method of claim 11 wherein either:
- a drain of the first switching FET and a drain of the second switching FET are coupled to a common node, the common node receiving the bias voltage; or
- a source of the first switching FET and a source of the second switching FET are coupled to a common node, the common node receiving the bias voltage.
13. The method of claim 8 wherein the measured value of the RF signal is a measured RF peak voltage of the RF signal.
14. The method of claim 13 wherein the bias voltage is varied such that, as the RF peak voltage increased, the bias voltage is decreased.
15. The method of claim 8 wherein the at least one switching FET receives the bias voltage at a drain of the at least one switching FET.
16. The method of claim 8 wherein the at least one switching FET receives the bias voltage at a source of the at least one switching FET.
17. The method of claim 8 further comprising a blocking inductor positioned between the bias voltage source and the switching circuits, the blocking inductor blocking an RF current from returning to the bias voltage source.
18. The method of claim 8 wherein each switching circuit further comprises at least one biasing FET positioned between the bias voltage source and the at least one switching FET, the biasing FET providing the voltage bias to the at least one switching FET.
19. A semiconductor processing tool comprising:
- a plasma chamber configured to deposit a material onto a substrate or etch a material from the substrate; and
- an impedance matching network operably coupled to the plasma chamber, the matching network comprising: an RF input operably coupled to an RF source providing an RF signal; an RF output operably coupled to the plasma chamber; and an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance of the EVC; wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor; wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET; and
- wherein the bias voltage is varied based on a measured value of the RF signal to keep an offset at or slightly above a desired bias voltage.
20. A method of fabricating a semiconductor, the method comprising:
- placing a substrate in a plasma chamber configured to deposit a material layer on the substrate or etch a material layer from the substrate;
- energizing plasma within the plasma chamber by providing an RF signal from an RF source to the plasma chamber to perform the deposition or etching; and
- while energizing the plasma, carrying out an impedance match by an impedance matching network coupled between the plasma chamber and the RF source, the matching network comprising: an RF input operably coupled to the RF source; an RF output operably coupled to the plasma chamber; and an electronically variable capacitor (EVC) comprising discrete capacitors, each discrete capacitor having a corresponding switching circuit for switching in and out the discrete capacitor to alter a total capacitance of the EVC; wherein each switching circuit comprises at least one switching field-effect transistor (FET) operably coupled to the corresponding discrete capacitor to cause the switching in and out of the discrete capacitor; wherein for each switching circuit, when the switching circuit is switched OFF to switch out the corresponding discrete capacitor, the at least one switching FET receives a bias voltage from a bias voltage source to reduce a capacitance variability of the at least one switching FET; and
- wherein the bias voltage is varied based on a measured value of the RF signal to keep an offset at or slightly above a desired bias voltage.
| 3828281 | August 1974 | Chambers |
| 4110700 | August 29, 1978 | Rosen et al. |
| 4679007 | July 7, 1987 | Reese et al. |
| 4692643 | September 8, 1987 | Tokunaga et al. |
| 4751408 | June 14, 1988 | Rambert |
| 4929855 | May 29, 1990 | Ezzeddine |
| 5012123 | April 30, 1991 | Ayasli et al. |
| 5079507 | January 7, 1992 | Ishida et al. |
| 5654679 | August 5, 1997 | Mavretic et al. |
| 5815047 | September 29, 1998 | Sorensen et al. |
| 5849136 | December 15, 1998 | Mintz et al. |
| 5880921 | March 9, 1999 | Tham et al. |
| 5889252 | March 30, 1999 | Williams et al. |
| 5971591 | October 26, 1999 | Vona et al. |
| 6046641 | April 4, 2000 | Chawla et al. |
| 6137367 | October 24, 2000 | Ezzedine et al. |
| 6252354 | June 26, 2001 | Collins et al. |
| 6400012 | June 4, 2002 | Miller et al. |
| 6424232 | July 23, 2002 | Mavretic et al. |
| 6583572 | June 24, 2003 | Veltrop et al. |
| 6621372 | September 16, 2003 | Kondo et al. |
| 6657395 | December 2, 2003 | Windhorn |
| 6677828 | January 13, 2004 | Harnett et al. |
| 6703080 | March 9, 2004 | Reyzelman et al. |
| 6791274 | September 14, 2004 | Hauer et al. |
| 6794951 | September 21, 2004 | Finley |
| 6818562 | November 16, 2004 | Todorow et al. |
| 6888313 | May 3, 2005 | Blackburn et al. |
| 6888396 | May 3, 2005 | Hajimiri et al. |
| 6946847 | September 20, 2005 | Nishimori et al. |
| 6967547 | November 22, 2005 | Pellegrini et al. |
| 7004107 | February 28, 2006 | Raoux et al. |
| RE39051 | March 28, 2006 | Harnett |
| 7071786 | July 4, 2006 | Inoue et al. |
| 7095178 | August 22, 2006 | Nakano et al. |
| 7113761 | September 26, 2006 | Bickham et al. |
| 7122965 | October 17, 2006 | Goodman |
| 7164236 | January 16, 2007 | Mitrovic et al. |
| 7199678 | April 3, 2007 | Matsuno |
| 7251121 | July 31, 2007 | Bhutta |
| 7298091 | November 20, 2007 | Pickard et al. |
| 7298128 | November 20, 2007 | Bhutta |
| 7304438 | December 4, 2007 | Kishinevsky |
| 7332981 | February 19, 2008 | Matsuno |
| 7439610 | October 21, 2008 | Weigand |
| 7480571 | January 20, 2009 | Howald et al. |
| 7495524 | February 24, 2009 | Omae et al. |
| 7498908 | March 3, 2009 | Gurov |
| 7514935 | April 7, 2009 | Pankratz |
| 7518466 | April 14, 2009 | Sorensen et al. |
| 7535312 | May 19, 2009 | McKinzie, III |
| 7602127 | October 13, 2009 | Coumou |
| 7642879 | January 5, 2010 | Matsuno |
| 7666464 | February 23, 2010 | Collins et al. |
| 7714676 | May 11, 2010 | McKinzie, III |
| 7728602 | June 1, 2010 | Valcore et al. |
| 7745955 | June 29, 2010 | Kirchmeier et al. |
| 7755300 | July 13, 2010 | Kishinevsky et al. |
| 7764140 | July 27, 2010 | Nagarkatti et al. |
| 7777567 | August 17, 2010 | Polizze |
| 7852170 | December 14, 2010 | McKinzie, III |
| 7863996 | January 4, 2011 | Cotter et al. |
| 7868556 | January 11, 2011 | Xia |
| 7872523 | January 18, 2011 | Sivakumar et al. |
| 7917104 | March 29, 2011 | Manssen et al. |
| 7969096 | June 28, 2011 | Chen |
| 8008982 | August 30, 2011 | McKinzie, III |
| 8040068 | October 18, 2011 | Coumou et al. |
| RE42917 | November 15, 2011 | Hauer et al. |
| 8089026 | January 3, 2012 | Sellers |
| 8102954 | January 24, 2012 | Coumou |
| 8110991 | February 7, 2012 | Coumou |
| 8203859 | June 19, 2012 | Omae et al. |
| 8217731 | July 10, 2012 | McKinzie, III |
| 8217732 | July 10, 2012 | McKinzie, III |
| 8228112 | July 24, 2012 | Reynolds |
| 8237501 | August 7, 2012 | Owen |
| 8264154 | September 11, 2012 | Banner et al. |
| 8278909 | October 2, 2012 | Fletcher |
| 8289029 | October 16, 2012 | Coumou |
| 8299867 | October 30, 2012 | McKinzie, III |
| 8314561 | November 20, 2012 | Fisk et al. |
| 8330432 | December 11, 2012 | Van Zyl et al. |
| 8334657 | December 18, 2012 | Xia |
| 8334700 | December 18, 2012 | Coumou et al. |
| 8335479 | December 18, 2012 | Koya et al. |
| 8344559 | January 1, 2013 | Van Zyl et al. |
| 8344801 | January 1, 2013 | Owen et al. |
| 8368308 | February 5, 2013 | Banna et al. |
| 8368469 | February 5, 2013 | Mohammadi et al. |
| 8395322 | March 12, 2013 | Coumou |
| 8416008 | April 9, 2013 | Van Zyl et al. |
| 8436643 | May 7, 2013 | Mason |
| 8461842 | June 11, 2013 | Thuringer et al. |
| 8466736 | June 18, 2013 | Reynolds |
| 8487706 | July 16, 2013 | Li et al. |
| 8502689 | August 6, 2013 | Chen et al. |
| 8513889 | August 20, 2013 | Zhang et al. |
| 8520413 | August 27, 2013 | Tran et al. |
| 8536636 | September 17, 2013 | Englekirk |
| 8552665 | October 8, 2013 | Larson et al. |
| 8558633 | October 15, 2013 | McKinzie, III |
| 8559907 | October 15, 2013 | Burgener et al. |
| 8564381 | October 22, 2013 | McKinzie |
| 8569842 | October 29, 2013 | Weis et al. |
| 8576010 | November 5, 2013 | Yanduru |
| 8576013 | November 5, 2013 | Coumou |
| 8587321 | November 19, 2013 | Chen et al. |
| 8620236 | December 31, 2013 | Manssen et al. |
| 8624501 | January 7, 2014 | Nagarkatti et al. |
| 8633782 | January 21, 2014 | Nagarkatti et al. |
| 8638159 | January 28, 2014 | Ranta et al. |
| 8649754 | February 11, 2014 | Burgener et al. |
| 8659335 | February 25, 2014 | Nagarkatti et al. |
| 8674606 | March 18, 2014 | Carter et al. |
| 8680928 | March 25, 2014 | Jeon et al. |
| 8686796 | April 1, 2014 | Presti |
| 8710926 | April 29, 2014 | Nagarkatti et al. |
| 8716984 | May 6, 2014 | Mueller et al. |
| 8723423 | May 13, 2014 | Hoffman et al. |
| 8742669 | June 3, 2014 | Carter et al. |
| 8773019 | July 8, 2014 | Coumou et al. |
| 8779859 | July 15, 2014 | Su et al. |
| 8781415 | July 15, 2014 | Coumou et al. |
| 8815329 | August 26, 2014 | Ilic et al. |
| 8847561 | September 30, 2014 | Karlieek et al. |
| 8884180 | November 11, 2014 | Ilic et al. |
| 8884525 | November 11, 2014 | Hoffman et al. |
| 8890537 | November 18, 2014 | Valcore, Jr. et al. |
| 8912835 | December 16, 2014 | Nagarkatti et al. |
| 8928329 | January 6, 2015 | Downing et al. |
| 9083343 | July 14, 2015 | Li et al. |
| 9190993 | November 17, 2015 | Li |
| 9306533 | April 5, 2016 | Anton |
| 10269540 | April 23, 2019 | Carter et al. |
| 20020060914 | May 23, 2002 | Porter et al. |
| 20030007372 | January 9, 2003 | Porter et al. |
| 20030046013 | March 6, 2003 | Gerrish |
| 20060170367 | August 3, 2006 | Bhutta |
| 20060198077 | September 7, 2006 | Bhutta |
| 20060232471 | October 19, 2006 | Coumou |
| 20070075784 | April 5, 2007 | Pettersson et al. |
| 20070139122 | June 21, 2007 | Nagarkatti et al. |
| 20080179948 | July 31, 2008 | Nagarkatti et al. |
| 20080180179 | July 31, 2008 | Polizzo |
| 20080197854 | August 21, 2008 | Valcore et al. |
| 20090207537 | August 20, 2009 | Coumou |
| 20100001796 | January 7, 2010 | Sivakumar et al. |
| 20100073104 | March 25, 2010 | Cotter et al. |
| 20100123502 | May 20, 2010 | Bhutta et al. |
| 20100194195 | August 5, 2010 | Coumou et al. |
| 20100201370 | August 12, 2010 | Coumou et al. |
| 20100231296 | September 16, 2010 | Nagarkatti et al. |
| 20110241781 | October 6, 2011 | Owen et al. |
| 20110247696 | October 13, 2011 | Zolock et al. |
| 20120013253 | January 19, 2012 | Coumou |
| 20120062322 | March 15, 2012 | Owen |
| 20120188007 | July 26, 2012 | Van Zyl et al. |
| 20120262064 | October 18, 2012 | Nagarkatti et al. |
| 20130043854 | February 21, 2013 | Tran et al. |
| 20130169359 | July 4, 2013 | Coumou |
| 20130193867 | August 1, 2013 | Van Zyl et al. |
| 20130207738 | August 15, 2013 | Mason |
| 20130222055 | August 29, 2013 | Coumou et al. |
| 20130257311 | October 3, 2013 | Tran et al. |
| 20130314163 | November 28, 2013 | Costa |
| 20130320853 | December 5, 2013 | Carter et al. |
| 20140009248 | January 9, 2014 | Granger-Jones |
| 20140028389 | January 30, 2014 | Coumou |
| 20140028398 | January 30, 2014 | Owen |
| 20140049250 | February 20, 2014 | Brown et al. |
| 20140055034 | February 27, 2014 | Coumou |
| 20140061156 | March 6, 2014 | Brouk et al. |
| 20140062303 | March 6, 2014 | Hoffman et al. |
| 20140097908 | April 10, 2014 | Fisk, II et al. |
| 20140117861 | May 1, 2014 | Finley et al. |
| 20140117872 | May 1, 2014 | Finley |
| 20140118031 | May 1, 2014 | Rughoonundon et al. |
| 20140210345 | July 31, 2014 | Hoffman |
| 20140210551 | July 31, 2014 | Mueller |
| 20140218076 | August 7, 2014 | Coumou et al. |
| 20140220913 | August 7, 2014 | Coumou et al. |
| 20140231243 | August 21, 2014 | Finley |
| 20140232266 | August 21, 2014 | Finley et al. |
| 20140266492 | September 18, 2014 | Radomski et al. |
| 20140306742 | October 16, 2014 | Menzer et al. |
| 20140320013 | October 30, 2014 | Coumou et al. |
| 20150115289 | April 30, 2015 | Fursin et al. |
| 20150171860 | June 18, 2015 | Blin |
| 20170345620 | November 30, 2017 | Coumou |
| 0840349 | May 1998 | EP |
| 0840350 | May 1998 | EP |
| 2006096589 | September 2006 | WO |
Type: Grant
Filed: Jan 6, 2020
Date of Patent: May 24, 2022
Patent Publication Number: 20200144032
Assignee:
Inventor: Michael Gilliam Ulrich (Delran, NJ)
Primary Examiner: Samuel S Outten
Application Number: 16/735,088
International Classification: H01J 37/32 (20060101); H03H 7/40 (20060101); H01L 21/67 (20060101); H01L 21/3213 (20060101); H01L 21/02 (20060101); H01L 21/285 (20060101); H01L 21/311 (20060101); C23C 16/505 (20060101);
